RefSeq annotation of the CHM13 genome

The RefSeq annotation of the CHM13 genome is available here.

This annotation was created by using the Liftoff program [1] (v1.6.3, with options -copies -sc 0.95 -polish -exclude_partial -chroms) to map across all human genes in RefSeq [2] annotation release 110 from the GRCh38.p14 genome to the CHM13v2.0 genome, a complete, gap-free human genome published by the Telomere-to-Telomere (T2T) consortium [3].

Of the 58,700 genes (19,871 protein coding) and 179,372 transcripts (130,361 protein coding) annotated on the main chromosomes and unplaced contigs in GRCh38, we successfully lifted over 58,272 genes (19,767 protein coding) and 178,688 transcripts (130,426 protein coding). We also used the "-copies" option in Liftoff to identify additional gene copies in CHM13, using a minimum sequence identity threshold of 95% at the DNA level. With this threshold, we found 2,393 additional gene copies (239 protein coding). Finally, we added the ribosomal DNA (rDNA) annotations on the acrocentric chromosomes from the CHM13v2.0 annotation, which were created using the Comparative Annotation Toolkit (CAT) [4] and Liftoff. These steps resulted in a total gene count of 61,322 (20,006 protein coding) and a total transcript count of 181,715 (130,426 protein coding). The table below shows the number of genes of each biotype in GRCh38 and CHM13.

Gene biotype Number of genes in GRCh38 Number of genes mapped onto CHM13
protein coding 19871 20006
lncRNA 17793 18389
pseudogene 15357 16030
miRNA 1914 2047
transcribed pseudogene 1221 1262
snoRNA 1195 1188
tRNA 431 522
V segment 239 245
V segment pseudogene 189 209
snRNA 153 192
J segment 98 79
ncRNA 51 49
rRNA 38 325
misc RNA 36 37
D segment 32 0
C region 21 23
antisense RNA 19 19
other 14 13
J segment pseudogene 7 6
C region pseudogene 5 5
Y RNA 4 7
vault RNA 4 4
scRNA 4 4
telomerase RNA 1 1
RNase P RNA 1 1
RNase MRP RNA 1 1
ncRNA pseudogene 1 1
Total 58700 60655

A comparison of the genes using gene IDs to match annotations is shown in the figure below. Genes were compared between RefSeq v110 on GRCh38, the Liftoff version of RefSeq on CHM13, and the NCBI version of RefSeq on CHM13. The NCBI version, shown in blue, was produced using the Gnomon pipeline, which only mapped approximately half of the protein coding genes.

Figure: Venn diagrams showing the overlaps between the RefSeq annotation of GRCh38 (yellow), the RefSeq annotation of CHM13 produced here by Liftoff (red), and the NCBI RefSeq annotation produced internally at NCBI (blue) using the Gnomon pipeline. The large orange regions show the near-complete overlap between our RefSeq annotation (Liftoff-based) and the GRCh38 annotation.

1. Shumate, A. & Salzberg, S. L. Liftoff: accurate mapping of gene annotations. Bioinformatics 37, 1639–1643 (2021).

2. O’Leary, N. A. et al. Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res 44, D733–D745 (2016).

3. Nurk, S. et al. The complete sequence of a human genome. Science (1979) 376, 44–53 (2022).

4. Fiddes, I. T. et al. Comparative Annotation Toolkit (CAT)-simultaneous clade and personal genome annotation. Genome Research 28, 1029–1038 (2018).